Abstract:Objective To construct a risk prediction model for perforation during endoscopic submucosal dissection（ESD）in early gastric cancer，and to provide a scientific basis for clinicians to assess surgical risks and formulate reasonable treatment plans. Methods Clinical data of 102 patients with early gastric cancer who underwent ESD treatment in the Third Affiliated Hospital of Soochow University from January to December 2024 were collected，including general information，tumor-related characteristics，and operation-related indicators. Univariate analysis and multivariate logistic regression analysis were used to identify related factors and independent risk factors，and a perforation risk prediction model was constructed. The predictive efficacy of the model was evaluated by receiver operating characteristic（ROC）curve. Results Among the 102 patients，there were 20 cases in the perforation group and 82 cases in the non-perforation group. Multivariate logistic regression analysis showed that tumor diameter >3 cm，tumor located in the upper part of the gastric body，and operation time >120 min were independent risk factors for ESD- related perforation in early gastric cancer（OR=3.557，95%CI：1.777-7.120，P<0.01；OR=2.892，95%CI：1.406-5.949，P=0.004；OR=4.276，95% CI：2.253- 8.117，P<0.01）. Hosmer-Lemeshow goodness of fit test Results showed that the regression model had a good fitting effect on the data（χ2=6.328，P=0.612）. The area under the ROC curve（AUC）for the combined factors in predicting the risk of ESD perforation was 0.86（95%CI：0.80-0.92），indicating good predictive accuracy. Conclusion The successfully constructed risk prediction model for ESD-related perforation in early gastric cancer incorporates independent risk factors such as tumor diameter，tumor location，and operation time. It can predict the perforation risk relatively accurately.